The rationale for this study stems from the observation that hip fractures are age-related, and are influenced by a number of risk factors. Since only a small proportion of people develop age-related fractures, it may be concluded that hip fractures are not inevitable. The identification of the circumstances under which hip fractures occur would provide an opportunity to understand the etiology of this condition and to formulate intervention measures. The investigators conducted several preliminary sutdies which led them to classify the conditions associated with hip fractures into five classes. Of particular significance is their Class D, which includes risk factors of hyperthyroidism, thyroidectomy, gonodal hypofunction, gastrectomy, hemiparesis, etc. In addition, a number of factors were considered protective against hip fractures such as urolithiasis, estrogen therapy, obesity, gout, or pseudogout (Class E). The study design involves the estimation of the incidence of hip fractures in specified cohorts of patients in comparison to sex, race, and age matched controls. In addition, comparisons will be made within each cohort in subgroups of patients with different characteristics. For each of the cohort studies, a team of investigators will have the responsibility for data collection and analysis, but the statistical design and overall analytic support will be provided by the core. After each of the cohorts is identified, such as cophorectomy, thyroidectomy, gastrointestinal, and surgery cohorts, pertinent information will be abstracted from the records. The computerized indices identify the potential case or control, but the actual data has to be collected manually by the nurse abstracter.